Attrition Rates in Dietary Behaviour Change Interventions for Women with Gestational Diabetes Mellitus: A Systematic Review and Meta-analysis

Abstract

Background: Dietary behaviour change interventions are known to be effective for women with gestational diabetes (GDM). However, more information is needed about the nature of such dietary interventions, including the behaviour change techniques used and intervention attrition rates. Aims: To identify behaviour change techniques employed within dietary interventions for women with GDM, obtain a pooled estimate of the attrition rate in such interventions, and investigate methodological or intervention characteristics that may influence attrition. Methods: A systematic search of six electronic databases identified studies for review and meta-analysis. The most common behaviour change technique clusters, identified deductively using the Behaviour Change Technique taxonomy, were 'Feedback and Monitoring', 'Shaping Knowledge' and 'Antecedents'. The primary meta-analysis of attrition using proportions as the effect size measure employed a random effect model. Publication bias and between-study heterogeneity were investigated and explored through mixed-effect moderator analyses and univariate meta-regression models. Results: The pooled attrition rate across 25 included studies (Nparticipants = 926) was 16.5% (95% CI [11.2, 23.61), increasing to 26.8% (95% CI [23.67, 30.11]) when adjusted for publication bias. The sample demonstrated high heterogeneity (Q = 120.733, p <001; 7 = 913; P= 80.12%), although exploratory analyses did not reveal significant moderators. Conclusions: Moderate attrition occurs in diet-only GDM interventions. Clear and consistent reporting of intervention designs is essential for implementing, replicating and synthesising effective biomedical and behavioural components of GDM dietary interventions. Keywords: Gestational Diabetes Mellitus; dietary intervention; healthy eating; behaviour change; intervention design; meta-analysis